International Journal on Future Revolution in Computer Science & Communication Engineering ISSN: 2454-4248

Volume: 4 Issue: 1 341 – 345

_______________________________________________________________________________________________
Self-Adaptive Multi-Population Jaya Algorithm based Reactive Power Reserve
Optimization Considering Voltage Stability Margin Constraints

Dr. Aditya Tiwary

Associate Professor
Electrical & Electronics Engineering Department
Institute of Engineering & Science
IPS Academy, Indore
Email ID: [email protected]

Abstract— In this paper an innovative technique is proposed to achieve the optimum value of reactive power reserve accounting voltage stability
margin constrains. Reactive power reserve and voltage stability are important issues for proper operation on the power system. This is achieved
by suitable settings of reactive power control variables. The fitness function has been minimizes using Self-adaptive multi-population based Jaya
Algorithm (SAMP - Jaya). The developed algorithm has been implemented on two IEEE test systems.
Keywords- Self-adaptive multi-population based Jaya Algorithm, voltage stability margin, reactive power reserve, reactive power control
variables
_________________________________________________*****_________________________________________________

PSO algorithm. Fuzzy technique to develop a reactive power

NOTATIONS optimization algorithm for hybrid system was developed by
QGp Reactive power output of pth generator buses Taghavi et al. [9]. Khazali et al. [10] applied harmony search
algorithm for obtaining optimal performance of the system
Upper and lower limit on reactive power
QGpmax , QGpmin output of pth bus
based on reactive power considerations. Genetic algorithm has
been employed for voltage stability margin enhancement and
Average reactive power output
QGpmax reactive power dispatch by Devaraj et al. [11]. Mousavi et al.
[12] developed a preventive strategy for reactive power
VSM Voltage stability margin
management along with VSM improvement. Singh et al. [13]
Threshold value of Voltage stability margin
VSM th developed a multi objective VAR management algorithm using
Upper and lower limits on load bus voltages modified differential evolution algorithm. Titare et al. [14] used
Vi max , Vi min voltage dependent reactive power reserves modeling for
UK Kth control variable voltage stability enhancement employing ensemble of mutation
and crossover strategies and parameters in differential
Lower and upper limits on control variable
U Kmin , U Kmax evolution (EPSDE). Fang et al. [15] developed a robust optimal
NC Number of total control variables reactive power reserves dispatch under stochastic environment
0 Current total load of load injected at buses employing chance constraints
P d relaxation – based method. Bhattacharya and Raj [16] used
modal analysis and L-index for optimization of reactive power
I. INTRODUCTION reserves based on differential evolution technique. Sun et al
Modern power system is highly complex in nature. In this [17] presented a bi- objective reactive power reserves
complex power system obtaining the desired voltage profile optimization algorithm to coordinate long and short term
along with voltage stability margin in very important and is a voltage stability considerations. Fang et al [18] developed an
matter of top priority. Since long proximity indicators have interval optimal reactive power reserve dispatch considering
been used for voltage security enhancement. Tiranuchit et al. uncertainties in the load and load direction. Rojar et al. [19]
[1] applied minimum singular value of jacobian to maintain presented an excellent review of various metaheuristic
desired voltage stability margin (VSM) and voltage profile. techniques used for optimal reactive reserves dispatch.
Sensitivity analysis has been used by Begovic et al. [2] for
improving voltage security. An algorithm for optimum In this paper an innovative methodology based on Self-
reactive power dispatch employing LP and an optimal adaptive multi-population based Jaya Algorithm (SAMP -
impedance solution on voltage stability index was developed Jaya) for reactive power reserve optimization is proposed. The
by Chebbo et al. [3]. Ajjarapu et al. [4] developed an optimal specific objectives are to limit reactive power generation on
planning strategy for reactive power against voltage instability both side i.e. over excitation as well as under excitation, such
employing repeated load flow runs up to voltage collapse point. that sufficient reactive power reserve is available on either side.
Bansilal et al. [5] proposed optimal reactive power dispatch It has been proposed to directly evaluate the VSM and it should
algorithm for voltage stability improvement. Kessel et al. [6] have at least some threshold. The proposed methodology has
estimated the voltage stability of the power system. Pande et al. been implemented on IEEE 14-bus and 30-bus standard test
[7] used functional link network for reactive power systems.
management and voltage stability enhancement. Titare et al [8]
developed an approach to mitigate probability of voltage
collapse accounting parameter uncertainties using improved
341
IJFRCSCE | January 2018, Available @ http://www.ijfrcsce.org
_______________________________________________________________________________________
International Journal on Future Revolution in Computer Science & Communication Engineering ISSN: 2454-4248
Volume: 4 Issue: 1 341 – 345
_______________________________________________________________________________________________
II. PROBLEM FORMULATION 4.1 Case-A: IEEE 14-Bus System [14]
The quadratic performance index is selected as fitness function Table-1 shows reactive power control variables (PV-bus
as proposed by Purey et al. [20] voltages, shunt compensations and OLTCs) and all load bus
voltages under base case condition. Table-2 shows the
QGp  QGpavg
F   p 1[ comparison of each algorithm to find the best optimal
NG
]2 (1)
QG max
p  QG min
p
control variable settings with and without optimization
using the proposed SAMP – Jaya optimization algorithm
Now QGpavg is given as follows [20]: and the results has been compared with Jaya [20], TLBO
[20], DE [20] and CAPSO [20] techniques. Table-3 shows
QPpavg  0.5[QGpmax  QGpmin ] (2) the comparison of reactive reserves at different generator
bus (bus nos. 1st, 2nd & 3rd) using SAMP - Jaya optimization
The objective function is minimized subject to following algorithm. Table-4 shows the comparison of SAMP - Jaya
constrains [20]: with other techniques reported in literature based on
(i) VSM  VSM th arithmetic mean value, standard deviation, best value, worst
value, frequency of convergence, standard error, length of
(ii) Vi min  Vi  Vi max confidence interval and confidence interval of fitness
(iii) U Kmin  U K  U Kmax K=1,2,3,4,….NC function. Fig. 1 shows a plot for the convergence of fitness
function with respect to number of iteration for SAMP -
(iv) All load bus voltages within limits at pd0 Jaya technique. It is observed that SAMP - Jaya gives much
better global optimal results than other optimization
and also (1  VSM Th
d
) Pd0 techniques reported in literature [20].

4.2 Case-B: IEEE 30-Bus System [22]

III. SELF-ADAPTIVE MULTI-POPULATION BASED JAYA
Table-5 shows reactive power control variables and all load
ALGORITHM (SAMP - JAYA): AN OVERVIEW:
bus voltages under base case condition. Table-6 shows the
The self-adaptive multi-population based Jaya algorithm for best optimal control variable settings with and without
solving the constrained and unconstrained numerical and optimization. Table-7 shows the comparison of reactive
engineering optimization problem was proposed by Rao et al. reserves at different generator bus (bus nos. 1st, 2nd, 5th, 8th,
[21]. It is based on the concept that the solution obtained for a 11th & 13th) using SAMP - Jaya technique. Table-8 shows the
given problem should move towards the best solution and comparison of proposed SAMP – Jaya optimization algorithm
avoid the worst solution. First, the initial population is with Jaya [20], TLBO [20], DE [20] and CAPSO [20]
generated having population size (NP) and number of design techniques based on arithmetic mean value, standard
variables (D) is decided. Now divide the total population into deviation, best value, worst value, frequency of convergence,
m sub populations depending on quality of solution. Identify standard error, length of confidence interval and confidence
the best and worst solution in each and every sub population. interval of objective function. Fig. 2 shows a plot for
Obtain the new value as follows: convergence of fitness function with respect to number of
Yï ' '  Yi  rand1 (Ybest,i  Yi )  rand2 (Yworst,i  Yi ) iteration for SAMP – Jaya. It is observed that SAMP - Jaya
(3) gives much better global optimal results than in comparison
Where with Jaya [20], TLBO [20], DE [20] and CAPSO [20]
Yi is previous value techniques.
Ybest,i is best solution v. CONCLUSION
An innovative methodology has been presented for reactive
Yworst,i is worst solution power reserves. The objective is to maintain voltage profile
and voltage stability margin by using SAMP – Jaya
rand1,rand2 are random numbers having range of [0,1] optimization algorithm. The developed algorithm performance
Accept the better solution in each sub population. Merge the is been analysed based on different criteria such as mean
entire sub population together. Now check if previous best value, median value, mean deviation, variance, standard
solution of entire population is better than the current best deviation, best value, worst value, frequency of convergence,
solution in the entire population. If yes, then m is decreased by standard error, length of confidence interval, confidence
1, else m is increased by 1. The procedure is terminated if a interval, class interval & proportionate frequencies of fitness
maximum number of generations have been executed. function. The results obtained by the proposed SAMP – Jaya
optimization algorithm has given better results than in
IV. RESULTS AND DISCUSSIONS: comparison with the different methods shown in the literature.
In this paper, Self-adaptive multi-population based Jaya
Algorithm (SAMP - Jaya) has been applied to obtain
optimum reactive power reserve using reactive power REFERENCES
control variables such as PV- bus voltages, OLTC and shunt [1] A. Tiranuchit, R. J. Thomas, “A posturing strategy against
compensations on IEEE 14-bus and 30-bus standard test voltage instabilities in electric power systems”, IEEE Trans. on
systems. PS, 1988, Vol. 3, No.1, pp.81-93.

342
IJFRCSCE | January 2018, Available @ http://www.ijfrcsce.org
_______________________________________________________________________________________
International Journal on Future Revolution in Computer Science & Communication Engineering ISSN: 2454-4248
Volume: 4 Issue: 1 341 – 345
_______________________________________________________________________________________________
[2] M. M. Begovic, A. G. Phadke, “Control of voltage stability stability margin”, Int. J. of Engineering trends and technology,
using sensitivity analysis”, IEEE Trans. on PS, 1992, Vol. 7, Vol. 51, 2017, pp. 106-14.
No.1 , pp. 114-123. [21] R. V. Rao, A. Saroj, “A self-adaptive multi-population based
[3] A. M. Chebbo, M. R. Irving, M. J. H. Sterling, “Reactive power Jaya algorithm for engineering optimization”, Swarm and
dispatch incorporating voltage stability”, Proc. IEEE Part-C, Evolutionary computation, 2017,
1992, Vol. 139, No. 3, pp.253-260. http://dx.doi.org/10.1016/j.swevo.2017.04.008.
[4] V. Ajjarapu, P. L. Lau, S. Battula, “An optimal reactive power [22] L. D. Arya, A. Koshti, “Anticipatory load shedding for line
planning strategy against voltage collapse”, IEEE Trans on PS, overload alleviation using Teaching learning based optimization
1994, Vol. 9, No. 2, pp.906-917. (TLBO)’s”, International Journal of Electrical Power & Energy
[5] Bansilal, D. Thukaram, K. Parthsarthy, “optimal reactive power Systems, Vol. 63, 2014, pp. 862-877.
dispatch algorithm for voltage stability improvement”, Int. J. of
Electrical Power and Energy systems, Vol. 18, No. 7, 1996 Table-1. Load flow solution for 14-bus test system under stressed
pp.461-468. condition. [20]
[6] P. Kessel, H. Glavitsch, “Estimating the voltage stability of Total load (Sd)=3.6758pu, Static voltage stability
power systems”, IEEE Trans. on PWRD, Vol. 1, No. 3, pp. 346- limit=4.6858pu
354. S. Control Control Load bus Load bus
[7] V. S. Pande, L. D. Arya, S. C. Choube, “Artificial neural No. variables variables voltages voltage
network based reactive reserve management for voltage stability magnitude magnitude (pu)
enhancement”, Proc. Int. Conf. on Power systems, Kathmandu, (pu)
Nepal, Dec.3-6, 2004. 1 V1 1.0812 V4 0.8248
[8] L. S. Titare, L. D. Arya, “An approach to mitigate voltage 2 V2 1.0485 V5 0.8618
collapse accounting uncertainties using improved particle swarm 3 V3 1.0739 V6 0.9522
optimization”, Applied soft Computing Journal, Vol. 19, No. 4, 4 BSH4 0.0015 V7 0.8618
sept.2009, pp.1197-1207. 5 BSH12 0.0057 V8 0.9696
[9] R. Taghavi, A. R. Seifi, M. N. Pourahmadi, “Fuzzy reactive
6 TAP4 1.0657 V9 0.8291
power optimization in hybrid power systems”, Int. J Electrical
7 TAP10 1.0673 V10 0.8126
power Energy systems, Vol. 42, No. 1, 2012, pp 375-383.
V11 0.8114
[10] A. H. Khazali, M. Kalantar, “optimal reactive power dispatch
based on harmony search algorithm”, Int. J Electrical Power V12 0.7970
Energy Systems, Vol. 33, No. 3, 2011, pp.684-692. V13 0.7917
[11] D. Devraj, J. P. Roselyn, “Genetic algorithm based reactive V14 0.7897
power dispatch for voltage stability improvement”, Int. J.
Electrical power Energy system, Vol. 32, No. 10, 2010, pp.1151-
1156. Table-2 Reactive power control variables using SAMP - Jaya algorithms for
[12] O. A. Mousavi, M. Bozorg, R. Cherkaowi, “Preventive reactive IEEE 14-bus system (Sdt)=3.6758pu.
power management for improving voltage stability margin”, S. Reactive Base SAMP JAYA TLBO DE [20] CAPS
No. control case – [20] [20] O [20]
Electric Power system Research, Vol. 96, 2013, pp. 36-46.
variable [20] JAYA
[13] H. Singh, L. Shrivastava , “Modified differential evolution 1 Tap4 1.0657 0.9315 0.9317 0.9320 0.9326 0.9284
algorithm for multi objective VAR management”, Int. J. of 2 Tap10 1.0673 0.9268 0.9266 0.9254 0.9258 0.9217
Electrical Power Energy systems, Vol. 55, 2014 , pp. 731-740. 3 Qc4 0.0015 0.0512 0.0508 0.0370 0.0447 0.0409
[14] L. S. Titare, P. Singh, L. D. Arya, S. C. Choube, “Optimal 4 Qc12 0.0057 0.0478 0.0473 0.0483 0.0357 0.0318
reactive power rescheduling based EPSDE algorithm to enhance 5 V1 1.0812 1.0790 1.0788 1.0798 1.0797 1.0776
static voltage stability”, Int. J. of Electrical Power and Energy 6 V2 1.0485 1.0428 1.0428 1.0445 1.0457 1.0447
systems, Vol. 63, 2014, pp. 588-599. 7 V3 1.0739 1.0695 1.0693 1.0704 1.0716 1.0693
[15] S. Fang, H. Cheng, G. Xu, Q. Zhou, H. He, P. Zeng,
“Stochastic reactive power dispatch considering voltage areas”, Table-3 Reactive power reserve at generator buses and fitness function using
Int. Trans. on Electrical Energy system, Vol. 27, No. 3, March SAMP – Jaya technique for IEEE 14-bus system (Sdt)=3.6758pu.
2017, Doi 10.10.02/etep.2269. S. Methodology Reactive power Reserve (pu) Total Fitness
[16] B. Bhattacharya, S. Raj, “Differential evolution technique for the No. Qgk(res)1 Qgk(res)1 Qgk(res)1 reactive function
optimization of reactive power reserves”, Journal of circuits power
reserve
systems and computers (world’s scientific), Vol. 26, No. 10,
(pu)
Oct. 2017, pp.1750-155. 1 SAMP – Jaya 2.6988 0.8185 0.014 3.5318 0.2082
[17] Q. Sun, H. Cheng, Y. Song, “Bi-objective reactive power 2 JAYA [20] 2.6991 0.8181 0.014 3.5312 0.2088
reserve optimization to co- ordinate long and short term voltage 3 TLBO [20] 2.7114 0.7916 0.0152 3.5183 0.221
stability”, IEEE Access, Date of publication 9, May 2017, DOI: 4 DE [20] 2.737 0.7646 0.012 3.5136 0.2326
10.1109/ ACCESS. 2017.2701826. 5 CAPSO [20] 2.7628 0.707 0.0319 3.5016 0.2489
[18] S. Fang, H. Cheng, Y. Song, Z. Ma, Z. Zhu, P. Zeng, L. Yao, 6 Base Case [20] 2.5295 0.6628 0.0398 3.2321 0.3381
“Interval optimal reactive power dispatch considering generator
rescheduling”, IET Generation, Transmission and distribution ,
Vol. 10, No. 8, May 2016, pp. 1833-1841.
[19] D. G. Rojar, J. L. Lezma, W. Villa, “Meta heuristic Techniques
Applied to the optimal reactive power dispatch : a review”,
IEEE Latin America Transactions, Vol. 14, No. 5, May 2016,
pp. 2253-2263.
[20] P. Purey, R. Arya, “Application of Jaya algorithm for reactive
power reserve optimization accounting constrains on voltage

343
IJFRCSCE | January 2018, Available @ http://www.ijfrcsce.org
_______________________________________________________________________________________
International Journal on Future Revolution in Computer Science & Communication Engineering ISSN: 2454-4248
Volume: 4 Issue: 1 341 – 345
_______________________________________________________________________________________________
Table -4 statistical inferences based on proposed SAMP – Jaya Table-6 Reactive power control variables using SAMP - Jaya algorithms for
techniques for IEEE 14-bus system. IEEE 30-bus system (Sdt)=4.6759pu.
Optimization SAMP JAYA TLBO DE CAPSO [20] S. Control Base SAMP JAYA TLBO DE CAPSO
methods – Jaya [20] [20] [20] No. variable case – [20] [20] [20] [20]
Arithmetic 0.2102 0.2132 0.2275 0.2430 0.2669 [20] JAYA
mean value 1 Tap11 1.0686 0.9246 0.9247 0.9232 0.9232 0.9253
of OF 2 Tap12 1.0693 1.0260 1.0263 1.0238 1.0238 1.0275
Median value 0.2078 0.2115 0.2255 0.2421 0.2653 3 Tap15 1.0563 0.9305 0.9314 0.9327 0.9349 0.9266
of OF 4 Tap36 0.9215 1.0743 1.0759 1.0839 1.0692 1.0791
Mean 2.00E- 2.00E- 4.00E- 1.50E- 1.50E-05 5 Qc10 0.0106 0.1753 0.1750 0.1756 0.1543 0.1556
deviation of 05 05 05 05 6 Qc24 0.0040 0.0378 0.0380 0.0372 0.0356 0.0375
OF 7 V1 1.0842 1.0821 1.0820 1.0768 1.0833 1.0710
Variance of 2.38E- 2.47E- 5.05E- 8.03E- 1.77E-04 8 V2 1.0476 1.0315 1.0317 1.0266 1.0352 1.0194
OF 05 05 05 05 9 V5 1.0112 1.0098 1.0097 1.0011 1.0111 0.9980
Standard 0.0045 0.0049 0.0071 0.0089 0.0133 10 V8 1.0262 1.0141 1.0141 1.0140 1.0261 1.0234
deviation of 11 V11 1.0845 1.0839 1.0838 1.0833 1.0846 1.0768
OF 12 V13 1.0928 1.0912 1.0912 1.0807 1.0921 1.0875
Best value of 0.2056 0.2088 0.2210 0.2326 0.2489
OF
Worst value 0.2245 0.2273 0.2480 0.2662 0.2951 Table-7 Reactive power reserve at generator buses and fitness function using
of OF SAMP – Jaya technique for IEEE 30-bus system (Sdt)=4.6759pu.
Frequency of 14 13 11 10 10 Met Reactive power Reserve (pu) Tota Fitne
convergence hod Qg Qgk(r Qgk Qgk Qgk Qgk l ss
Confidence 0.95 0.95 0.95 0.95 0.95 olog k(re es)2 (res) (res) (res) (res) reac funct
level y s)1 5 8 11 13 tive ion
Determined 2.0452 2.0452 2.0452 2.0452 2.0452 pow
value for the er
Engg. rese
Application rve
Standard 0.0020 0.0023 0.0032 0.0041 0.0061 (pu)
error of the SA 0.9 0.222 0.03 0.30 0.05 0.03 1.62 1.46
mean OF MP 753 8 25 28 45 01 78 38
Confidence 0.2101    0.02147 0.2243    0.02307
.2109    0.2155 .2389    0.2471
2608    0.2730 –
interval of Jaya
the OF JAY 0.9 0.222 0.03 0.30 0.05 0.03 1.62 1.46
Length of 0.0085 0.0094 0.0131 0.0167 0.0249 A 759 8 37 76 47 13 60 92
confidence [20]
interval of TL 0.9 0.209 0.08 0.22 0.04 0.05 1.61 1.51
the OF BO 908 9 04 92 41 89 33 19
OF-objective function [20]
Table-5. Load flow solution for 30-bus test system under stressed condition. DE 1.0 0.202 0.07 0.18 0.06 0.03 1.60 1.54
[20] [20] 301 7 91 97 84 8 8 09
Total load (Sd)=4.6759pu, Static voltage stability limit=6.2231pu CA 0.9 0.367 0.08 0.03 0.07 0.03 1.58 1.58
S. Control Control Load Load bus PSO 931 7 55 03 16 97 79 97
No. variables variables bus voltage [20]
magnitude voltages magnitude Bas 1.2 0.227 0.07 0.09 - - 1.31 2.29
(pu) (pu) e 278 2 29 65 0.07 0.23 47 25
1 V1 1.0842 V3 1.0231 Cas 49 48
2 V2 1.0476 V4 1.0105 e
3 V5 1.0112 V6 1.0052 [20]
4 V8 1.0262 V7 0.9902
5 V11 1.0845 V9 0.9400
6 V13 1.0928 V10 0.8948
7 BSH10 0.0106 V12 0.9516
8 BSH24 0.0040 V14 0.9135
9 TAP11 1.0686 V15 0.8996
10 TAP12 1.0693 V16 0.9110
11 TAP15 1.0563 V17 0.8873
12 TAP36 0.9215 V18 0.8686
V19 0.8578
V20 0.8651
V21 0.8674
V22 0.8693
V23 0.8703
V24 0.8511
V25 0.8593
V26 0.8379
V27 0.8749
V28 0.9981
V29 0.8311
V30 0.8084

344
IJFRCSCE | January 2018, Available @ http://www.ijfrcsce.org
_______________________________________________________________________________________
International Journal on Future Revolution in Computer Science & Communication Engineering ISSN: 2454-4248
Volume: 4 Issue: 1 341 – 345
_______________________________________________________________________________________________
Table -8 statistical inferences based on proposed SAMP – Jaya techniques for
IEEE 30-bus system.
Optimization SAMP JAYA TLBO DE CAPS
methods – Jaya [20] [20] [20] O [20]
Arithmetic 1.4805 1.4814 1.5348 1.5781 1.6461
mean value of
OF
Median value 1.4768 1.4787 1.5298 1.5751 1.6476
of OF
Mean deviation 2.00E- 2.00E- 4.50E- 5.00E- -5.00E-
of OF 05 05 05 05 05
Variance of OF 1.40E- 1.40E- 4.02E- 7.64E- 1.38E-
04 04 04 04 03
Standard 0.0115 0.0118 0.0200 0.0276 0.0371
deviation of OF
Best value of 1.4678 1.4692 1.5119 1.5409 1.5897
OF
Worst value of 1.5105 1.5142 1.5849 1.6479 1.7177
OF
Frequency of 13 12 11 10 9
Fig. 1. Plot of convergence of fitness function with respect to number
convergence of iteration using SAMP – Jaya techniques for IEEE 14-bus system.
Confidence 0.95 0.95 0.95 0.95 0.95
level
Determined 2.0452 2.0452 2.0452 2.0452 2.0452
value for the
Engg.
Application
Standard error 0.0041 0.0054 0.0091 0.0126 0.0169
of the mean OF
Confidence 1.4758  1.4760  1
1.4866 .5257  1.5655
1.4868  1.6292
1.5439    1.6630
1.5907
interval of the
OF
Length of 0.0210 0.0221 0.0372 0.0515 0.0691
confidence
interval of the
OF
OF-objective function

Fig. 2. Plot of convergence of fitness function with respect to number

of iteration using SAMP – Jaya techniques for IEEE 30-bus system.

345
IJFRCSCE | January 2018, Available @ http://www.ijfrcsce.org
_______________________________________________________________________________________